Efficient low-temperature thermochemical CO2 splitting enabled by Gibbs free energy engineering

Abstract

Thermochemical CO₂ splitting presents a promising mass solution to the global energy crisis and climate change. Due to the chemical inertness of CO₂, its conversion to CO through thermochemical processes without reductants is thermodynamically disfavored, particularly at lower reaction temperatures coupled with high reaction rates. In this study, we developed novel perovskite-based oxides using an integrated approach that combines high-throughput computational screening and machine learning. By systematically modulating cation types, concentrations, and oxygen non-stoichiometry in perovskites, perovskite structures were constructed based on the modified tolerance factor. Through establishing quantitative relationships between perovskite configurations and Gibbs free energy, we identified 8451 promising candidates with practical potential from an initial dataset of 279 142 perovskite combinations. Experimental validation confirmed the feasibility of this strategy, wherein the La_0.5Sm_0.125Sr_0.375Co_0.25Fe_0.125Ti_0.625O₃ perovskite oxide achieved a record CO yield of 1.834 mmol g⁻¹ at 1100 °C, which is 250 °C lower than the reduction temperature of conventional thermochemical processes. The synergy between computational design and experimental validation establishes a generalizable framework for developing high-activity redox materials and offers a viable solution for low-temperature thermochemical CO₂ splitting.